Inverse associations between soy and prostate cancer and the contribution of hormones to prostate cancer prompted the current study to determine whether soy protein could alter serum hormones in men. Thirty-five men consumed milk protein isolate (MPI), low-isoflavone soy protein isolate (SPI) (low-iso SPI; 1.64 +/- 0.19 mg isoflavones/d), and high-iso SPI (61.7 +/- 7.35 mg isoflavones/d) for 57 d each in a randomized crossover design. Twenty-four-hour urine samples indicated that urinary isoflavones were significantly increased by the high-iso SPI relative to the low-iso SPI and MPI. Serum collected on d 1, 29, and 57 of each treatment revealed that dihydrotestosterone (DHT) and DHT/testosterone were significantly decreased by the low-iso SPI [9.4% (P = 0.036) and 9.0% (P = 0.004), respectively] and the high-iso SPI [15% (P = 0.047) and 14% (P = 0.013), respectively], compared with the MPI at d 57. Other significant effects included a decrease in testosterone by the low-iso SPI relative to the MPI (P = 0.023) and high-iso SPI (P = 0.020) at d 29; an increase in dehydroepiandrosterone sulfate by the low-iso SPI relative to the MPI at d 29 (P = 0.001) and relative to the MPI (P = 0.0003) and high-iso SPI (P = 0.005) at d 57; and increases in estradiol and estrone by the low-iso SPI relative to the MPI at d 57 (P = 0.010 and P = 0.005, respectively). In conclusion, soy protein, regardless of isoflavone content, decreased DHT and DHT/testosterone with minor effects on other hormones, providing evidence for some effects of soy protein on hormones. The relevance of the magnitude of these effects to future prostate cancer risk requires further investigation.